Compliance
NVIDIA NIM in Brazil
Structured issue-spotting for deploying NVIDIA NIM in Brazil, read against the rules for that jurisdiction.
- Source
- Rules engine over a generic brief — no anchored quote on this page
- Verified
- Evidence not verified
- Confidence
- Low
Structured issue-spotting to support your own review — not legal advice. Verify against the cited primary sources and your counsel.
01What this reading assumes
- Jurisdiction
- Brazil
- Principal framework
- Lei nº 13.709/2018 (LGPD) is the general data protection statute. The AI-relevant provisions are article 20 on review of decisions taken solely on automated processing, articles 33 to 36 on international transfer bases and adequacy criteria, article 38 letting the authority demand a data protection impact report at any time, and article 46 requiring security measures applied from the design phase of the product through to its execution. Commencement was staged: the authority’s own provisions from December 2018 and administrative sanctions from August 2021. Resolution 15/2024 sets incident notification at three business days to both the regulator and the affected data subjects, and there is still no regulation specific to impact reports — the authority points controllers at the high-risk definition in its small-processing-agent regulation instead.
- Delivery assessed
- Self-hosted
- Data leaves the network
- no
- Vendor home jurisdiction
- United States
- Verified vendor positions
- none — every vendor position below is a question, not an assurance
- Rules evaluated
- 37
- Rules fired
- 10
Assumptions about use
- An internal deployment used by employees, not a public-facing product.
- A person reads the output before acting on it — but that is not recorded, so the engine reports it as a gap rather than assuming it.
- No significant automated decision is taken about a person by the system alone.
02Issues to work through
Brazil
Three business days, to the regulator and to the people affected
Brazil’s security incident notification regulation sets a three-business-day clock and it runs to affected data subjects as well as to the regulator, with an incident register to be kept. Teams routinely plan for a regulator notification and discover the subject notification at the point they can least afford to draft it, which is why the templates belong in the runbook before launch.
- Required checks
- Draft the data-subject notification template before an incident, not during one.
- Decide who decides an incident is notifiable, and make sure they can be reached inside three business days.
- Set up the incident register now; reconstructing one afterwards is not possible.
- Vendor questions
- How quickly will you notify us of an incident affecting our data, and does that fit inside a three-business-day clock that starts with our own awareness?
- Technical controls
- Alert on the signals that would make an AI incident visible — anomalous retrieval volumes, prompt-injection indicators, unexpected egress.
Human review required — take this to your counsel
Brazil’s AI bill passed the Senate and is still waiting in the Chamber
The Chamber of Deputies’ own record shows PL 2338/2023 awaiting the rapporteur’s opinion in the special committee constituted to consider it. It passed the Senate in December 2024 and has not been voted by the Chamber. There is no Brazilian AI act, no risk classification to complete and no conformity assessment — and the regulator is nonetheless active.
- Required checks
- Do not build a Brazilian AI risk classification; there is nothing to classify against.
- Read the bill anyway if the deployment has a long life, because its treatment of automated decisions is where the regulator is already heading.
The regulator is already working on AI, and publishing what it looks at
The data protection authority published the first results of its AI regulatory sandbox in July 2026, having supervised companies testing AI systems in a controlled environment focused on governance, security, transparency and anonymisation. Nothing in it binds anybody. It is the clearest available signal of what the authority will expect when it does issue AI rules, and it is free to read.
- Required checks
- Read the published sandbox material before designing the governance artefacts, rather than inventing your own categories.
- Note that the authority’s regulatory agenda targets the automated-decision article specifically.
Cross Cutting
Self-hosting moves the security obligation to you
Keeping data on your own hardware answers the transfer question and creates an operations question. Patching, backups, key management, monitoring and incident response are now yours, and an unpatched inference server on the office network is a worse outcome than a well-run vendor.
- Required checks
- Name the person responsible for patching each component, and the cadence.
- Confirm backups exist, are encrypted, and have been restored at least once.
- Confirm there is an incident response path that includes this system.
- Technical controls
- Encrypt at rest and in transit, including between the application and the inference server.
- Centralise authentication through the existing identity provider rather than local accounts.
- Keep an audit log of who queried what, and protect it from the people it records.
- Subscribe to security advisories for each component and track upgrade lag.
Being able to reconstruct a decision months later
The question that arrives after a complaint is what the system was shown and what it produced on a particular day. Models change, prompts change, and indexes are rebuilt, so the answer has to be recorded at the time. Without it, the only available response is that the output cannot be reproduced.
- Required checks
- Decide what is recorded per interaction: model and version, prompt template version, retrieved document ids, output, reviewer and outcome.
- Set how long those records are kept, balanced against the retention duties that also apply to them.
- Vendor questions
- Does the vendor pin model versions, and how much notice is given before a model is retired or changed?
- Technical controls
- Version prompt templates in source control and log the version used.
- Log the model identifier and version returned by the provider, not the one you requested.
Confidentiality duties bind independently of data protection law
Material can be entirely free of personal data and still be the material a contract stops you disclosing. Client retainers, non-disclosure agreements, supplier contracts and common-law duties are the usual sources, and several of them require consent before a third party processes the material at all — which a model API call is.
- Required checks
- Review the confidentiality clauses in the contracts covering the material going into the system.
- Identify any contract requiring notice or consent before a subcontractor processes the material.
- Decide whether the deployment needs a confidentiality carve-out negotiated into new contracts.
- Vendor questions
- Will the vendor accept a confidentiality undertaking beyond its standard terms?
- Which staff at the vendor can access customer content, under what controls?
- Technical controls
- Segregate the most sensitive corpora into an index that the general assistant cannot reach.
Human review required — take this to your counsel
We were not told whether a person reviews the output
Where output influences a decision about a person, the reviewer has to be able to disagree with it. That needs three things a rubber-stamp review lacks: enough information to judge, enough time to judge, and an override that is used often enough to be real. Design it before the volume makes it impossible.
- Required checks
- Name the role that reviews the output and what they see when they do.
- Decide what evidence is retained about each review, so the practice can be shown to exist.
- Set a threshold below which the system must not act without review.
- Vendor questions
- Does the product expose the retrieved context and the confidence behind a suggestion, or only the answer?
- Technical controls
- Show the reviewer the retrieved sources next to the suggestion, not the suggestion alone.
- Record the reviewer’s decision, including overrides, as part of the audit trail.
Human review required — take this to your counsel
An AI deployment creates new copies of the data
Vector indexes, prompt logs, completion caches, evaluation datasets, fine-tuning checkpoints and backups are all copies of the source material in places the existing retention schedule does not mention. Deletion requests are the moment this is discovered, because deleting the source document does not delete its embedding.
- Required checks
- List every store the deployment creates and add each to the retention schedule.
- Establish how a deletion request propagates to the index, the caches and the logs.
- Establish how long backups keep material that has been deleted from the live system.
- Vendor questions
- What does the vendor retain, where, and for how long after we delete our copy?
- Technical controls
- Store the source document id with every embedding so deletion can cascade.
- Set time-to-live on prompt and completion logs rather than relying on manual cleanup.
Who can reach the model, the index and the weights
A self-hosted stack has three access surfaces that are easy to leave open: the inference endpoint, the vector index, and the weights on disk. Retrieval also carries an authorisation problem an ordinary application does not have — the index must not return a document to someone who could not open it in the source system.
- Required checks
- Confirm the inference endpoint is not reachable from outside the network and requires authentication.
- Confirm retrieval filters by the requesting user’s permissions, not only by relevance.
- Confirm who can read the model files and the index volume at the operating-system level.
- Technical controls
- Bind the inference server to a private interface and put an authenticating proxy in front of it.
- Carry document-level access control into the index and enforce it at query time.
- Encrypt the volume holding the weights and the index, and restrict it to the service account.
- Rotate API keys and keep them out of client-side code and container images.
What ends up in a prompt, and where it goes next
Even with inference inside the network, prompts and retrieved context accumulate in logs, traces and caches, and system prompts can often be extracted from the output. The leak path is internal rather than external, but it is still a copy of the source material in a new place.
- Required checks
- Write down which categories of information may be entered into a prompt, and tell people.
- Establish what the system prompt contains and whether disclosing it would matter.
- Establish which shadow tools staff are already using; the policy has to name the permitted ones.
- Vendor questions
- Are prompts and completions retained, for how long, and can retention be set to zero?
- Are prompts used for abuse monitoring, and if so who can read them and for how long?
- Technical controls
- Redact or block high-risk patterns before the prompt leaves the application.
- Keep prompt and completion logs out of general-purpose observability tools.
- Set an explicit retention period on prompt logs and enforce it.
Structured issue-spotting to support your own review — not legal advice. Verify against the cited primary sources and your counsel.
03What this reading does not know
- Whether any of the data falls into a special or sensitive category.
- Whether any material is covered by legal professional privilege.
- Whether a person reviews the output before it is acted on.
04Instruments these issues point at
- proposalPL 2338/2023Brazil’s risk-based general AI bill, covering rights, transparency, human review, governance and penalties. It passed the Senate and is still awaiting the rapporteur’s opinion in the Chamber of Deputies’ special committee. It is not law and has not been voted by the Chamber.
- regulationANPD Res. 15/2024The security incident notification regulation under the LGPD. It sets who must be told, in what form and how fast — three business days to notify both the regulator and the affected data subjects — and requires an incident register to be kept.
- regulationANPD Res. 2/2022An eased LGPD regime for small processing agents. Its definition of high-risk processing is the ANPD’s stated interim benchmark for deciding when a data protection impact report is needed, because no regulation specific to impact reports has been issued.
- guidanceANPD AI SandboxAn ANPD pilot supervising a small number of companies testing AI systems in a controlled regulatory environment, focused on governance, security, transparency and anonymisation. It is the clearest signal of what the regulator will expect when it does issue AI rules.
05Vendor documents being watched
- Data processing agreementhttps://www.nvidia.com/en-us/agreements/data-processing-addendum/nvidia-cloud-services-data-processing-addendum/not yet fetched
- Privacy policyhttps://www.nvidia.com/en-us/about-nvidia/privacy-policy/not yet fetched
- Security pagehttps://www.nvidia.com/privacy-center/not yet fetched
- Terms of servicehttps://developer.nvidia.com/legal/termsnot yet fetched
06Ask about your own deployment
This page reads the rules against a generic organisation. Your size, industry, data and existing contracts change which of these issues matter and which fall away.